A dynamic monitoring and analysis system for mine roof pressure and rock burst

Abstract

The present invention discloses a dynamic monitoring and analysis system for mine roof pressure and rock burst pressure; it comprises a multifunctional sensor, a data acquisition module, a data transmission module, a data analysis and processing platform, a digital twin model, an early warning module, and a power supply module; the multifunctional sensor of the present invention adopts multi-point deployment and distributed monitoring to realize the collection of roof pressure, roof displacement, rock burst pressure fluctuation, rock burst impact force and gas concentration and composition inside the tunnel; the data twin model generates a model of the tunnel according to the collected data, which is convenient for generating a visual interface for the tunnel, and further provides data-driven decision support; the data analysis and processing platform and the early warning module analyze and process the collected data, predict the changes in roof pressure or rock burst pressure, and perform probability calculation on the analysis results through multinomial logistic regression to determine whether there are safety hazards, issue an alarm signal, and even automatically start protective measures to ensure the safety of the tunnel.

Description

Technical Field

[0001] The present invention belongs to the technical field of tunnel detection, and more specifically, particularly relates to a dynamic monitoring and analysis system for mine roof pressure and rock burst pressure. Background Art

[0002] The roof pressure and rock burst problems in the mining process have always been important factors affecting mine safety and miners' life safety. With the continuous increase in mining depth and the increasingly complex mining environment, traditional roof pressure monitoring and rock burst prediction methods have gradually exposed many shortcomings in dealing with the complex pressure environment of deep mines.

[0003] Mine roof pressure mainly refers to the pressure on the roof during mining due to the combined effects of mining operations, rock movement, groundwater and other factors. Excessive or uneven distribution of roof pressure may cause serious disasters such as roof collapse and rock crushing, threatening the safety of the mine.

[0004] Rock burst refers to the violent dynamic effect caused by the sudden release of rock mass stress during the mining process, which is usually manifested as strong vibration and instantaneous rock fracture. This phenomenon not only causes damage to equipment, but also poses a great threat to the lives of miners.

[0005] At present, roof pressure monitoring in mines mainly relies on sensor equipment such as pressure sensors and displacement meters to monitor the changes in roof pressure in real time. These monitoring methods can provide a certain degree of early warning and data support, but in mines with complex geological conditions and large mining depths, it is often difficult to accurately reflect the dynamic changes of roof pressure, and it is difficult to predict sudden changes and abnormalities in roof pressure.

[0006] The monitoring and early warning of rock burst mainly rely on seismic wave monitoring technology, stress sensors and numerical simulation. Although these methods can monitor the occurrence of rock burst to a certain extent, due to the suddenness and strong dynamic characteristics of rock burst, traditional monitoring technology is often difficult to achieve timely and accurate early warning, especially in high stress areas and complex structural areas, the monitoring effect is not ideal.

[0007] However, there are some problems with the existing technology: the existing detection equipment cannot transmit data in real time when in use, which easily causes high data delays, and cannot visualize the conditions in the tunnels, and cannot predict the probability of accidents based on the collected data, and cannot provide early warning. Therefore, we propose a dynamic monitoring and analysis system for mine roof pressure and rock burst pressure. Summary of the invention

[0008] In view of the problems existing in the prior art, the purpose of the present invention is to provide a dynamic monitoring and analysis system for mine roof pressure and impact rock pressure. A digital twin model is generated through the collected data, and accident probability prediction is performed based on the data, thereby realizing visualization of the situation inside the tunnel, facilitating observation of the tunnel, and predicting accident probability based on the processed data to achieve early warning.

[0009] To achieve the above-mentioned object, the present invention provides the following technical solutions: a dynamic monitoring and analysis system for roof pressure and rock burst for mining, comprising a multifunctional sensor, a data acquisition module, a data transmission module, a data analysis and processing platform, a digital twin model, an early warning module, and a power supply module;

[0010] The multifunctional sensor is used to collect the roof pressure and rock burst data in the tunnel in real time. The multifunctional sensor integrates a pressure sensor, a displacement sensor, a rock burst sensor, a stress sensor and a gas sensor. The pressure sensor is used to monitor the pressure change of the roof; the displacement sensor is used to measure the displacement and deformation of the roof; the rock burst sensor is used to monitor the fluctuation of underground rock burst; the stress sensor is used to measure the impact force, vibration or stress change generated when the rock burst is abnormal; the gas sensor is used to detect the gas concentration and composition inside the tunnel;

[0011] The data acquisition module is used to connect with the multifunctional sensor and acquire the data collected by the multifunctional sensor, and a plurality of the data acquisition modules are provided, each of which is electrically connected to a plurality of the multifunctional sensors. After acquiring the data from the data acquisition module, the data acquisition module pre-processes the data;

[0012] The data transmission module is used to transmit the data collected by the data acquisition module, and transmit the data collected by the data acquisition module to the data analysis and processing platform;

[0013] The data analysis and processing platform is used to process, analyze and visualize data through deep learning and generate corresponding monitoring reports;

[0014] The digital twin model is used to construct a tunnel model with the collected data, which is used to reflect the status and behavior of the physical objects, processes, systems or environments in the simulated tunnel in real time; by combining the multifunctional sensor, data acquisition technology, computer modeling and analysis, the digital twin model reproduces the real-time dynamic changes of the physical entity in the virtual environment, thereby providing data-driven decision support;

[0015] The early warning module performs probability calculation on the analysis results through multiple logistic regression, determines whether there are safety hazards, issues an alarm signal, and even automatically initiates protective measures;

[0016] The power supply module is used to supply power to the system and ensure stable operation of the system.

[0017] Specifically, the multifunctional sensor and the data acquisition module adopt multi-point deployment and distributed monitoring. According to the width and depth of the tunnel, monitoring points are deployed in multiple working faces or different areas to obtain ground pressure and pressure data at different locations in real time, provide more comprehensive monitoring, and can monitor the roof pressure and impact ground pressure of the tunnel in real time.

[0018] Specifically, the data acquisition module preprocesses the data, and the preprocessing includes data filtering, anomaly detection, missing value filling, data amplification and format unification;

[0019] The data filtering uses windows of different scales to perform median filtering, and then performs weighted fusion on the filtering results of different scales;

[0020]

[0021] Among them, y i represents the output data after filtering; S represents the number of scales, that is, the number of different scales used when performing the median filtering operation, and each scale corresponds to a filtering window of different sizes; w s Represents the weight of scale S, which is used to perform weighted averaging of filtering results of different scales; represents the result of median filtering the data at scale S. Specifically, It is all the data within the window range at scale S, that is, the neighborhood data at this scale; window s Represents the neighborhood window corresponding to the scale S. At different scales, the size of the neighborhood window is different.

[0022] Specifically, the calculation formula for the anomaly detection is as follows:

[0023] The Z-score of the data is obtained by calculating the difference between the data and the mean and dividing it by the standard deviation;

[0024]

[0025] Among them, x is the value of the data point, μ is the mean of the data set, σ is the standard deviation of the data set, and points with a value of Z greater than 3 or less than -3 are considered outliers and are removed;

[0026] The missing value compensation is used to fill the vacancies of outliers, and the missing values ​​are filled by the mean, and the mean is calculated as follows:

[0027]

[0028] Among them, x′ i is the value of non-missing data, and n is the number of non-missing data.

[0029] Specifically, the data amplification is used to adjust the data in proportion, and the amplification method adopted is minimum-maximum scaling;

[0030]

[0031] Among them, x i is the input data, x min is the minimum value in the data, x max is the maximum value in the data, x scalcd It is the scaled data, which is scaled to the range of [0, 1];

[0032] The format unification is used to unify the data scales or numerical forms of different features to facilitate subsequent processing;

[0033] The format uniformly adopts norm normalization:

[0034]

[0035] X is the data vector; ||X||2 is the L2 norm of vector X, i.e., the Euclidean distance,

[0036]

[0037] X normalized is the normalized data vector.

[0038] Specifically, the data transmission module adopts 5G network and wireless Mesh network;

[0039] The 5G network is used to realize remote data transmission of data collected by the multifunctional sensor in the lane, and realize real-time, multi-point, large-scale data transmission;

[0040] The wireless mesh network is used to establish a self-organizing network among multiple nodes in the lane, so as to enable the data transmission module to collect data information of the data collection module, and facilitate the data collection module to obtain the data collected by the multi-functional sensor.

[0041] Specifically, the calculation method of deep learning of the data analysis and processing platform is as follows:

[0042]

[0043] Among them: Al-1 is the output of the previous layer. For the input layer, A 0 is the input data matrix, and the data matrix includes roof pressure, roof displacement and deformation, rock burst pressure fluctuation, rock burst force, gas concentration and composition; W l is the weight matrix of the lth layer; b l is the bias vector; σ l is the activation function, Z l It is an important variable in the calculation process of a certain layer in a deep neural network, usually referring to the input of a linear combination;

[0044] The final output layer A l The corresponding value is the predicted roof pressure or rock burst pressure;

[0045] σ l The calculation formula of the activation function is as follows:

[0046]

[0047] Among them, z is the linear combination input to the Sigmoid function, which is usually composed of the activation value of the previous layer and the weight and bias of the current layer; e is the base of the natural logarithm, which is set to 2.718.

[0048] Specifically, the calculation of the digital twin model is as follows:

[0049] The digital twin model uses Moor's model to analyze the stress and deformation of the roadway as follows:

[0050]

[0051] Where: r is the radial stress; P0 is the initial formation stress; R is the tunnel radius; r is the distance from the tunnel center;

[0052] The hoop stress is calculated as follows:

[0053]

[0054] Used to calculate the stress distribution and possible deformation inside and outside the tunnel, helping to evaluate the stability of the tunnel;

[0055] Ground settlement and earthquake activity are calculated using foundation settlement theory:

[0056]

[0057] Where: S is the settlement; q is the soil bearing capacity; E is the elastic modulus of the soil; Area is the area where the load acts;

[0058] Temperature, humidity and gas concentration are calculated using the diffusion equation:

[0059]

[0060] Where: C is the gas concentration; D is the gas diffusion coefficient; is the Laplace operator of gas concentration, representing the spatial variation of concentration; Represents the change of concentration C with time t.

[0061] Specifically, the digital twin model is based on real-time data to update the model, which requires the use of some data-driven methods, such as Kalman filtering, to estimate the system state and update the model;

[0062] The dynamic update calculation of the digital twin model is as follows:

[0063] Prediction steps:

[0064]

[0065] Update steps:

[0066]

[0067] in, is the estimated system state; P kk-1 is the predicted error covariance; K k is the Kalman gain;

[0068] R is the observation noise covariance; z k is the actual observed value.

[0069] Specifically, the calculation formula of the multinomial logistic regression is as follows:

[0070]

[0071] Among them, z n is the linear model output corresponding to category n; N is the total number of categories; P(y=n|x) is the probability of predicting category N.

[0072] Technical effects and advantages of the present invention:

[0073] When the present invention is in use, the roof pressure, roof displacement, rock pressure fluctuation, rock pressure impact force and gas concentration and composition inside the tunnel are realized through multi-functional sensors, and the multi-functional sensors, data acquisition modules and data transmission modules are arranged at multiple points and distributed monitoring. According to the width and depth of the tunnel, monitoring points are arranged at multiple working faces or different areas to obtain the ground pressure and pressure data at different positions in real time, providing more comprehensive monitoring, and being able to monitor the roof pressure and rock pressure of the tunnel in real time;

[0074] The data twin model generates a model of the tunnel based on the collected data, which facilitates the generation of a visual interface for the tunnel, and is used to reflect the status and behavior of the physical objects, processes, systems or environments in the simulated tunnel in real time. The digital twin model reproduces the real-time dynamic changes of the physical entity in the virtual environment, thereby providing data-driven decision support.

[0075] The data analysis and processing platform and the early warning module analyze and process the collected data, and finally predict and calculate the situation inside the tunnel through the output layer value which is the predicted roof pressure or rock burst pressure change. The early warning module performs probability calculation on the analysis results through multiple logistic regression to determine whether there are safety hazards, issue an alarm signal, and even automatically initiate protective measures to ensure the safety of the tunnel.

[0076] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a schematic diagram of the system structure provided by the present invention. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0079] like Figure 1 As shown, a mining roof pressure and rock burst dynamic monitoring and analysis system provided by an embodiment of the present invention includes a multifunctional sensor, a data acquisition module, a data transmission module, a data analysis and processing platform, a digital twin model, an early warning module, and a power supply module;

[0080] The multifunctional sensor is used to collect the roof pressure and rock burst data in the tunnel in real time. The multifunctional sensor integrates pressure sensor, displacement sensor, rock burst sensor, stress sensor and gas sensor. The pressure sensor is used to monitor the pressure change of the roof; the displacement sensor is used to measure the displacement and deformation of the roof; the rock burst sensor is used to monitor the underground rock burst fluctuation; the stress sensor is used to measure the impact force, vibration or stress change generated when the rock burst is abnormal; the gas sensor is used to detect the gas concentration and composition inside the tunnel;

[0081] The data acquisition module is used to connect with the multifunctional sensor and acquire the data collected by the multifunctional sensor, and is provided with a plurality of data acquisition modules, each of which is electrically connected to a plurality of multifunctional sensors. After acquiring the data from the data acquisition module, the data acquisition module pre-processes the data;

[0082] The data transmission module is used to transmit the data collected by the data acquisition module, and transmit the data collected by the data acquisition module to the data analysis and processing platform;

[0083] The data analysis and processing platform is used to process, analyze and visualize data through deep learning and generate corresponding monitoring reports;

[0084] The digital twin model is used to construct a tunnel model based on the collected data, which is used to reflect the status and behavior of physical objects, processes, systems or environments in the simulated tunnel in real time. By combining multifunctional sensors, data acquisition technology, computer modeling and analysis, the digital twin model reproduces the real-time dynamic changes of the physical entity in the virtual environment, thereby providing data-driven decision support.

[0085] The early warning module uses multiple logistic regression to calculate the probability of the analysis results, determine whether there are safety hazards, issue an alarm signal, and even automatically initiate protective measures;

[0086] The power module is used to supply power to the system and ensure stable operation of the system.

[0087] In this embodiment, preferably, the multifunctional sensor and data acquisition module adopt multi-point deployment and distributed monitoring. According to the width and depth of the tunnel, monitoring points are deployed on multiple working faces or different areas to obtain ground pressure and pressure data at different locations in real time, providing more comprehensive monitoring, and being able to monitor the roof pressure and rock burst pressure of the tunnel in real time;

[0088] It should be noted that through multi-point deployment and distributed detection, a comprehensive inspection of the interior of the tunnel can be carried out, and the roof pressure and impact ground pressure of the tunnel can be monitored in real time.

[0089] In this embodiment, preferably, the data acquisition module preprocesses the data, and the preprocessing includes data filtering, anomaly detection, missing value filling, data amplification and format unification;

[0090] Data filtering uses windows of different scales for median filtering, and then performs weighted fusion on the filtering results of different scales;

[0091]

[0092] Among them, y irepresents the output data after filtering; S represents the number of scales, that is, the number of different scales used when performing the median filtering operation, and each scale corresponds to a filtering window of different sizes; w s Represents the weight of scale S, which is used to perform weighted averaging of filtering results of different scales; represents the result of median filtering the data at scale S. Specifically, It is all the data within the window range at scale S, that is, the neighborhood data at this scale; window s Represents the neighborhood window corresponding to the scale S. At different scales, the size of the neighborhood window is different;

[0093] It should be noted that data filtering is used to process data, eliminate noise in the data, and improve data accuracy. Weighted averaging is also used to facilitate weighted averaging of filtering results of different scales.

[0094] In this embodiment, preferably, the calculation formula for abnormality detection is as follows:

[0095] The Z-score of the data is obtained by calculating the difference between the data and the mean and dividing it by the standard deviation;

[0096]

[0097] Among them, x is the value of the data point, μ is the mean of the data set, σ is the standard deviation of the data set, and points with a value of Z greater than 3 or less than -3 are considered outliers and are removed;

[0098] Missing value compensation is used to fill in the gaps left by outliers, and the missing values ​​are filled by the mean, which is calculated as follows:

[0099]

[0100] Among them, x′ i is the value of non-missing data, n is the number of non-missing data;

[0101] It should be noted that by detecting outliers in the data, it is convenient to improve the accuracy of the data, and by filling in missing values ​​and outliers through the mean, the integrity of the data can be improved.

[0102] In this embodiment, preferably, data amplification is used to adjust the data in proportion, and the amplification method used is minimum-maximum scaling;

[0103]

[0104] Among them, x i is the input data, x minis the minimum value in the data, x max is the maximum value in the data, x scalcd It is the scaled data, which is scaled to the range of [0, 1];

[0105] Format unification is used to unify the data scale or numerical form of different features to facilitate subsequent processing;

[0106] The format uniformly adopts norm normalization:

[0107]

[0108] X is the data vector; ||X||2 is the L2 norm of vector X, i.e., the Euclidean distance,

[0109]

[0110] X normalized is the normalized data vector;

[0111] It should be noted that the data is amplified through minimum-maximum scaling to prevent data loss during transmission, or to prevent the data from being too weak and causing inaccurate data. At the same time, the data is scaled to the range of [0, 1] to improve the efficiency of the calculation and reduce the pressure of the calculation, and the data format is unified through norm normalization.

[0112] In this embodiment, preferably, the data transmission module adopts 5G network and wireless Mesh network;

[0113] The 5G network is used to achieve remote data transmission of data collected by multifunctional sensors in the lanes, realizing real-time, multi-point, and large-scale data transmission;

[0114] Drive the wireless mesh network to establish a multi-node self-organizing network in the lane, which is used to realize the data transmission module to collect data information from the data collection module, and facilitate the data collection module to obtain data collected by the multi-function sensor;

[0115] It should be noted that the 5G network can realize fast data transmission to prevent high delays during data transmission, and the setting of the wireless mesh network facilitates the centralization of data collected by distributed multi-functional sensors, thereby facilitating data concentration and transmission.

[0116] In this embodiment, preferably, the calculation method of deep learning of the data analysis and processing platform is as follows:

[0117]

[0118] Among them: A l-1is the output of the previous layer. For the input layer, A 0 is the input data matrix, and the data matrix includes roof pressure, roof displacement and deformation, rock burst pressure fluctuation, rock burst force, gas concentration and composition; W l is the weight matrix of the lth layer; b l is the bias vector; σ l is the activation function, Z l It is an important variable in the calculation process of a certain layer in a deep neural network, usually referring to the input of a linear combination;

[0119] The final output layer A l The corresponding value is the predicted roof pressure or rock burst pressure;

[0120] σ l The calculation formula of the activation function is as follows:

[0121]

[0122] Where z is the linear combination input to the Sigmoid function, usually composed of the activation value of the previous layer and the weight and bias of the current layer; e is the base of the natural logarithm, set to 2.718;

[0123] It should be noted that through deep learning, data analysis and calculation can be realized to obtain roof pressure or rock burst pressure, and sudden changes and anomalies in roof pressure can be calculated, as well as the occurrence of rock burst pressure can be monitored, which is convenient for improving mine production safety, protecting the lives of miners, and improving mine production efficiency.

[0124] In this embodiment, preferably, the calculation of the digital twin model is as follows:

[0125] The digital twin model uses Moor's model to analyze the stress and deformation of the roadway as follows:

[0126]

[0127] Where: r is the radial stress; P0 is the initial formation stress; R is the tunnel radius; r is the distance from the tunnel center;

[0128] The hoop stress is calculated as follows:

[0129]

[0130] Used to calculate the stress distribution and possible deformation inside and outside the tunnel, helping to evaluate the stability of the tunnel;

[0131] Ground settlement and earthquake activity are calculated using foundation settlement theory:

[0132]

[0133] Where: S is the settlement; q is the soil bearing capacity; E is the elastic modulus of the soil; Area is the area where the load acts;

[0134] Temperature, humidity and gas concentration are calculated using the diffusion equation:

[0135]

[0136] Where: C is the gas concentration; D is the gas diffusion coefficient; is the Laplace operator of gas concentration, representing the spatial variation of concentration; represents the change of concentration C with time t;

[0137] It should be noted that the environmental data in the tunnel (such as temperature, humidity, gas concentration, pressure, etc.) can be collected in real time through sensors and matched with the virtual model; through these data, the environment and equipment operating status in the tunnel can be monitored in real time, the changes in gas concentration at different locations can be simulated and compared with the actual sensor data.

[0138] In this embodiment, preferably, the digital twin model is based on real-time data to update the model, and some data-driven methods such as Kalman filtering are needed to estimate the system state and update the model;

[0139] The dynamic update calculation of the digital twin model is as follows:

[0140] Prediction steps:

[0141]

[0142] Update steps:

[0143]

[0144] in, is the estimated system state; P kk-1 is the predicted error covariance; K k is the Kalman gain;

[0145] R is the observation noise covariance; z k is the actual observed value;

[0146] It should be noted that the dynamic update of digital twins means that when the physical entity changes or new data is generated, the digital model responds in real time and updates its status, so that the performance of the virtual model is always consistent with the actual physical object; ensuring that the virtual world is synchronized with the real world, providing guarantees for real-time monitoring, remote diagnosis, fault prediction and decision support; through continuous updates, digital twins can reflect real-time changes in the real system, thereby enhancing prediction accuracy and foresight, and avoiding system performance degradation or failure; dynamic updates can provide engineers and managers with more accurate information to help with refined management and optimization of production processes.

[0147] In this embodiment, preferably, the calculation formula of multinomial logistic regression is as follows:

[0148]

[0149] Among them, z n is the linear model output corresponding to category n; N is the total number of categories; P(y=n|x) is the probability of predicting category N;

[0150] It should be noted that the calculation of multiple logistic regressions makes it easy to predict the probability of accidents occurring in the tunnel, to send out alarm signals, and even to automatically initiate protective measures.

[0151] The specific operation process of the present invention is as follows:

[0152] When in use, the power module inside the tunnel supplies power to the system to ensure stable operation of the system. Then, data is collected from the tunnel through multi-point deployment and distributed monitoring of multi-functional sensors. Monitoring points are deployed on multiple working faces or in different areas according to the width and depth of the tunnel to obtain ground pressure and pressure data at different locations in real time, providing more comprehensive monitoring and capable of real-time monitoring of the tunnel roof pressure and impact ground pressure. The multi-functional sensor is used to collect real-time roof pressure and impact ground pressure data in the tunnel. The multi-functional sensor integrates pressure sensors, displacement sensors, ground pressure sensors, stress sensors and gas sensors. The pressure sensor is used to monitor pressure changes on the roof. The multi-functional sensor is then connected through the data acquisition module to obtain the data collected by the multi-functional sensor. After obtaining data from the data acquisition module, the data acquisition module pre-processes the data; the processed data is transmitted through the data transmission module, and the data collected by the data acquisition module is transmitted to the data analysis and processing platform; the data analysis and processing platform is used to process, analyze and visualize the data through deep learning, and generate corresponding monitoring reports; and the collected data is used to construct a tunnel model through the digital twin model to reflect the status and behavior of physical objects, processes, systems or environments in the simulated tunnel in real time; the real-time dynamic changes of the entity are reproduced in the virtual environment, and then data-driven decision support is provided; finally, the analysis results are probabilistically calculated through the multi-logistic regression of the early warning module to determine whether there are safety hazards, issue an alarm signal, and even automatically start protective measures.

[0153] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic monitoring and analysis system for roof pressure and rock burst for mining, characterized in that: It includes multifunctional sensors, data acquisition modules, data transmission modules, data analysis and processing platforms, digital twin models, early warning modules, and power modules; The multifunctional sensor is used to collect the roof pressure and rock burst data in the tunnel in real time. The multifunctional sensor integrates a pressure sensor, a displacement sensor, a rock burst sensor, a stress sensor and a gas sensor. The pressure sensor is used to monitor the pressure change of the roof; the displacement sensor is used to measure the displacement and deformation of the roof; the rock burst sensor is used to monitor the fluctuation of underground rock burst; the stress sensor is used to measure the impact force, vibration or stress change generated when the rock burst is abnormal; the gas sensor is used to detect the gas concentration and composition inside the tunnel; The data acquisition module is used to connect with the multifunctional sensor and acquire the data collected by the multifunctional sensor, and a plurality of the data acquisition modules are provided, each of which is electrically connected to a plurality of the multifunctional sensors. After acquiring the data from the data acquisition module, the data acquisition module pre-processes the data; The data transmission module is used to transmit the data collected by the data acquisition module, and transmit the data collected by the data acquisition module to the data analysis and processing platform; The data analysis and processing platform is used to process, analyze and visualize data through deep learning and generate corresponding monitoring reports; The digital twin model is used to construct a tunnel model with the collected data, which is used to reflect the status and behavior of the physical objects, processes, systems or environments in the simulated tunnel in real time; by combining the multifunctional sensor, data acquisition technology, computer modeling and analysis, the digital twin model reproduces the real-time dynamic changes of the physical entity in the virtual environment, thereby providing data-driven decision support; The early warning module performs probability calculation on the analysis results through multiple logistic regression, determines whether there are safety hazards, issues an alarm signal, and even automatically initiates protective measures; The power supply module is used to supply power to the system and ensure stable operation of the system.

2. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The multifunctional sensor and the data acquisition module adopt multi-point deployment and distributed monitoring. Monitoring points are deployed on multiple working faces or in different areas according to the width and depth of the tunnel, and ground pressure and pressure data at different locations are acquired in real time, providing more comprehensive monitoring, and capable of real-time monitoring of the roof pressure and impact ground pressure of the tunnel.

3. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The data acquisition module preprocesses the data, and the preprocessing includes data filtering, anomaly detection, missing value filling, data amplification and format unification; The data filtering uses windows of different scales to perform median filtering, and then performs weighted fusion on the filtering results of different scales; Among them, y i represents the output data after filtering; S represents the number of scales, that is, the number of different scales used when performing the median filtering operation, and each scale corresponds to a filtering window of different sizes; w s Represents the weight of the scale S, which is used to perform weighted averaging of filtering results of different scales; represents the result of median filtering the data at scale S. Specifically, It is all the data within the window range at scale S, that is, the neighborhood data at this scale; window s Represents the neighborhood window corresponding to the scale S. At different scales, the size of the neighborhood window is different.

4. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 3, characterized in that: The calculation formula for the anomaly detection is as follows: The Z-score of the data is obtained by calculating the difference between the data and the mean and dividing it by the standard deviation; Among them, x is the value of the data point, μ is the mean of the data set, σ is the standard deviation of the data set, and points with a value of Z greater than 3 or less than -3 are considered outliers and are removed; The missing value compensation is used to fill the vacancies of outliers, and the missing values ​​are filled by the mean, and the mean is calculated as follows: Among them, x i is the value of non-missing data, and n is the number of non-missing data.

5. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 3, characterized in that: The data amplification is used to adjust the data in proportion, and the amplification method adopted is minimum-maximum scaling; Among them, x i is the input data, x min is the minimum value in the data, x max is the maximum value in the data, x scalcd It is the scaled data, which is scaled to the range of [0, 1]; The format unification is used to unify the data scales or numerical forms of different features to facilitate subsequent processing; The format uniformly adopts norm normalization: X is the data vector; ||X||2 is the L2 norm of vector X, i.e., the Euclidean distance, X normalized is the normalized data vector.

6. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The data transmission module adopts 5G network and wireless Mesh network; The 5G network is used to realize remote data transmission of data collected by the multifunctional sensor in the lane, and realize real-time, multi-point, large-scale data transmission; The wireless mesh network is used to establish a self-organizing network among multiple nodes in the lane, so as to enable the data transmission module to collect data information of the data collection module, and facilitate the data collection module to obtain the data collected by the multi-functional sensor.

7. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The calculation method of deep learning of the data analysis and processing platform is as follows: Among them: A l-1 is the output of the previous layer. For the input layer, A 0 is the input data matrix, and the data matrix includes roof pressure, roof displacement and deformation, rock burst pressure fluctuation, rock burst force, gas concentration and composition; W l is the weight matrix of the lth layer; b l is the bias vector; σ l is the activation function, Z l It is an important variable in the calculation process of a certain layer in a deep neural network, usually referring to the input of a linear combination; The final output layer A l The corresponding value is the predicted roof pressure or rock burst pressure; σ l The calculation formula of the activation function is as follows: Among them, z is the linear combination input to the Sigmoid function, which is usually composed of the activation value of the previous layer and the weight and bias of the current layer; e is the base of the natural logarithm, which is set to 2.

718.

8. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The calculation of the digital twin model is as follows: The digital twin model uses Moor's model to analyze the stress and deformation of the roadway as follows: Where: r is the radial stress; P0 is the initial formation stress; R is the tunnel radius; r is the distance from the tunnel center; The hoop stress is calculated as follows: Used to calculate the stress distribution and possible deformation inside and outside the tunnel, helping to evaluate the stability of the tunnel; Ground settlement and earthquake activity are calculated using foundation settlement theory: Where: S is the settlement; q is the soil bearing capacity; E is the elastic modulus of the soil; Area is the area where the load acts; Temperature, humidity and gas concentration are calculated using the diffusion equation: Where: C is the gas concentration; D is the gas diffusion coefficient; is the Laplace operator of gas concentration, representing the spatial variation of concentration; Represents the change of concentration C with time t.

9. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 8, characterized in that: The digital twin model is based on real-time data to update the model, which requires the use of some data-driven methods, such as Kalman filtering, to estimate the system state and update the model; The dynamic update calculation of the digital twin model is as follows: Prediction steps: Update steps: in, is the estimated system state; P kk-1 is the predicted error covariance; K k is the Kalman gain; R is the observation noise covariance; z k is the actual observed value.

10. A mining roof pressure and rock burst dynamic monitoring and analysis system according to claim 1, characterized in that: The calculation formula of the multinomial logistic regression is as follows: Among them, z n is the linear model output corresponding to category n; N is the total number of categories; P(y=n|x) is the probability of predicting category N.

Images